Automation gets described as magic, which is exactly why people are nervous about handing anything to it. It is not magic. It is a list of steps with a decision in the middle, and here it is running.
Change the numbers
The threshold is the only real decision in this flow. Put the request above it and a person gets involved; put it below and the flow finishes on its own. That single condition is what most people mean when they say they want something automated.
Notice that the flow does not decide anything on the finance branch. It routes, and then it waits for a human. Automation is usually about removing the carrying and the chasing, not the judgement.
Then tick the failure box
This is the part worth insisting on when somebody builds you a flow. A step that can fail will eventually fail: the system it writes to will be down, a password will expire, a field will get renamed. What matters is not whether it can break, it is whether anybody finds out.
The failure path here retries, gives up, and then tells a named person with the request attached. Nothing is silently lost. A flow without that last step is the expensive kind of broken, because it looks exactly like a flow that is working.
That failure mode has its own article: automation that breaks quietly is worse than none. And if you are weighing this against hiring, the two are not competing purchases.
What this is not showing
A real flow has more steps than six, and most of them are unglamorous: checking the request is not a duplicate, handling the case where the requester has left, deciding what happens when two approvals arrive at once. The shape is the same. The work is in the cases nobody mentions in the first conversation.
If you have a process that looks like this and you are not sure whether it is worth automating, describe what happens step by step, including what happens when something goes wrong. That last part usually decides it.